Integration of an automatic recommendation component into a digital library of Phd dissertations from the university of Novi Sad

Abstract

This paper presents the development of a recommendation system component, and its integration into the PHD UNS digital library - a Digital Library consisting of PhD Dissertations from the University of Novi Sad. This component is used to improve the PHD UNS user experience by providing a list of PhD dissertations recommended specifically for that user, and also re-ranking search results based on personalized recommendation scores. The developed component has been put in operational mode on: http://www.cris.uns.ac.rs/searchDissertations.jsf. Recommendation is performed on the basis of the user’s download history.A sub-system that is able to capture users’ explicit and implicit feedback has also been developed. This feedback will be used to evaluate whether the personalisation component improves PHD UNS users’ experience.